bioRxiv preprint doi: https://doi.org/10.1101/020107; this version posted August 6, 2015. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. 1

1 The evolution, diversity and host associations of rhabdoviruses 2 3 Ben Longdon1*, Gemma GR Murray1, William J Palmer1, Jonathan P Day1, Darren J 4 Parker2, 3, John J Welch1, Darren J Obbard4 and Francis M Jiggins1. 5 6 1Department of Genetics 7 University of Cambridge 8 Cambridge 9 CB2 3EH 10 UK 11 12 2 School of Biology 13 University of St. Andrews 14 St. Andrews 15 KY19 9ST 16 UK 17 18 3Department of Biological and Environmental Science, 19 University of Jyväskylä, 20 Jyväskylä, 21 Finland 22 23 4Institute of Evolutionary Biology, and Centre for Immunity Infection and Evolution 24 University of Edinburgh 25 Edinburgh 26 EH9 3JT 27 UK 28 29 *corresponding author 30 email: [email protected] 31 phone: +441223333945 32 33 bioRxiv preprint doi: https://doi.org/10.1101/020107; this version posted August 6, 2015. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. 2

34 Abstract 35 36 Metagenomic studies are leading to the discovery of a hidden diversity of RNA viruses, 37 but new approaches are needed predict the host species these poorly characterised 38 viruses pose a risk to. The rhabdoviruses are a diverse family of RNA viruses that 39 includes important pathogens of humans, animals and plants. We have discovered the 40 sequences of 32 new rhabdoviruses through a combination of our own RNA sequencing 41 of insects and searching public sequence databases. Combining these with previously 42 known sequences we reconstructed the phylogeny of 195 rhabdovirus sequences 43 producing the most in depth analysis of the family to date. In most cases we know 44 nothing about the biology of the viruses beyond the host they were identified from, but 45 our dataset provides a powerful phylogenetic approach predict which are vector-borne 46 viruses and which are specific to vertebrates or arthropods. By reconstructing ancestral 47 and present host states we found that switches between major groups of hosts have 48 occurred rarely during rhabdovirus evolution. This allowed us to propose 76 new likely 49 vector-borne vertebrate viruses among viruses identified from vertebrates or biting 50 insects. Our analysis suggests it is likely there was a single origin of the plant viruses and 51 arthropod-borne vertebrate viruses while vertebrate-specific viruses arose at least 52 twice. There are also two large clades of viruses that infect insects, including the sigma 53 viruses, which are vertically transmitted. There are also few transitions between aquatic 54 and terrestrial ecosystems. Viruses also cluster together at a finer scale, with closely 55 related viruses tending to be found in closely related hosts. Our data therefore suggest 56 that throughout their evolution rhabdoviruses have occasionally made a long distance 57 host jump before spreading through related hosts in the same environment. This 58 approach offers a way to predict the most probable biology and key traits of newly 59 discovered viruses. 60 61 Keywords 62 Virus, Host shift, Arthropod, Insect, Rhabdoviridae, Mononegavirales 63 bioRxiv preprint doi: https://doi.org/10.1101/020107; this version posted August 6, 2015. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. 3

64 Introduction 65 66 RNA viruses are an abundant and diverse group of pathogens. In the past, viruses were 67 typically isolated from hosts displaying symptoms of infection, before being 68 characterized morphologically and then sequenced following PCR [1, 2]. PCR-based 69 detection of novel RNA viruses is problematic as there is no single conserved region of 70 the genome of viruses from a single family, let alone all RNA viruses. High throughput 71 next generation sequencing technology has revolutionized virus discovery, allowing 72 rapid detection and sequencing of divergent virus sequences simply by sequencing total 73 RNA from infected individuals [1, 2] 74 75 One particularly diverse family of RNA viruses is the Rhabdoviridae. Rhabdoviruses are 76 negative-sense single-stranded RNA viruses in the order Mononegavirales [3]. They 77 infect an extremely broad range of hosts and have been discovered in plants, fish, 78 mammals, reptiles and a broad range of insects and other arthropods [4]. The family 79 includes important pathogens of humans and livestock. Perhaps the most well-known is 80 rabies virus, which can infect a diverse array of mammals and causes a fatal infection 81 killing 59,000 people per year with an estimated economic cost of $8.6 billion (US) [5]. 82 Other rhabdoviruses such as vesicular stomatitis virus and bovine ephemeral fever virus 83 are important pathogens of domesticated animals, whilst others are pathogens of crops 84 [3]. 85 86 Arthropods play a key role in transmission of many rhabdoviruses. Many viruses found 87 in vertebrates have also been detected in arthropods, including sandflies, mosquitoes, 88 ticks and midges [6]. The rhabdoviruses that infect plants are also often transmitted by 89 arthropods [7] and the rhabdoviruses that infect fish have the potential to be vectored 90 by ectoparasitic copepod sea-lice [8, 9]. Insects are not just mechanical vectors as 91 rhabdoviruses replicate upon infection of insect vectors [7]. Other rhabdoviruses are 92 insect-specific. In particular, the sigma viruses are a clade of vertically transmitted 93 viruses that infect dipterans and are well-studied in Drosophila [10-12]. Recently, a 94 number of rhabdoviruses have been found to be associated with a wide array of insect 95 and other arthropod species, suggesting they may be common arthropod viruses [13, 96 14]. Furthermore, a number of arthropod genomes contain integrated endogenous viral 97 elements (EVEs) with similarity to rhabdoviruses, suggesting that these species have 98 been infected with rhabdoviruses [15-18]. 99 100 Here we aimed to uncover the diversity of the rhabdoviruses, and examine how they 101 have switched between different host taxa during their evolutionary history. Insects

102 infected with rhabdoviruses commonly become paralysed on exposure to CO2 [19-21]. 103 We exploited this fact to screen field collections of flies from several continents for novel 104 rhabdoviruses that were then sequenced using RNA-sequencing (RNA-seq). Additionally 105 we searched for rhabdovirus-like sequences in publicly available RNA-seq data. We 106 identified 32 novel rhabdovirus-like sequences from a wide array of invertebrates and 107 plants, and combined them with recently discovered viruses to produce the most 108 comprehensive phylogeny of the rhabdoviruses to date. For many of the viruses we do 109 not know their true host range, so we used the phylogeny to identify a large number of bioRxiv preprint doi: https://doi.org/10.1101/020107; this version posted August 6, 2015. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. 4

110 new likely vector-borne viruses and to reconstruct the evolutionary history of this 111 diverse group of viruses. 112 113 114 Methods 115 116 Discovery of new rhabdoviruses by RNA sequencing 117 118 Diptera (flies, mostly Drosophilidae) were collected in the field from Spain, USA, Kenya, 119 France, Ghana and the UK (Data S1: http://dx.doi.org/10.6084/m9.figshare.1425432). 120 Infection with rhabdoviruses can cause Drosophila and other insects to become

121 paralysed after exposure to CO2 [19-21], so we enriched our sample for infected

122 individuals by exposing them to CO2 at 12°C for 15 mins, only retaining individuals that 123 showed symptoms of paralysis 30mins later. We extracted RNA from 79 individual 124 insects (details in Data S1 http://dx.doi.org/10.6084/m9.figshare.1425432) using 125 Trizol reagent (Invitrogen) and combined the extracts into two pools (retaining non- 126 pooled individual RNA samples). RNA was then rRNA depleted with the Ribo-Zero Gold 127 kit (epicenter, USA) and used to construct Truseq total RNA libraries (Illumina). 128 Libraries were constructed and sequenced by BGI (Hong Kong) on an Illumina Hi-Seq 129 2500 (one lane, 100bp paired end reads, generating ~175 million reads). Sequences 130 were quality trimmed with Trimmomatic (v3); Illumina adapters were clipped, bases 131 were removed from the beginning and end of reads if quality dropped below a 132 threshold, sequences were trimmed if the average quality within a window fell below a 133 threshold and reads less than 20 base pairs in length were removed. We de novo 134 assembled the RNA-seq reads with Trinity (release 2013-02-25) using default settings 135 and jaccard clip option for high gene density. The assembly was then searched using 136 tblastn to identify rhabdovirus-like sequences, with known rhabdovirus coding 137 sequences as the query. Contigs with hits were then reciprocally compared to Genbank 138 cDNA and RefSeq nucleotide databases using tblastn and only retained if they hit a 139 virus-like sequence. Raw read data were deposited in the NCBI Sequence Read Archive 140 (SRP057824). Putative viral sequences have been submitted to Genbank (accessions in 141 Tables S1 and S2 http://dx.doi.org/10.6084/m9.figshare.1502665). 142 143 As the RNA-seq was performed on pooled samples, we assigned rhabdovirus sequences 144 to individual insects by PCR on individual samples RNA. cDNA was produced using 145 Promega GoScript Reverse Transcriptase and random-hexamer primers, and PCR 146 performed using primers designed using the rhabdovirus sequences. Infected host 147 species were identified by sequencing the mitochondrial gene COI. We were unable to 148 identify the host species of the virus from a Drosophila affinis sub-group species 149 (sequences appear similar to both Drosophila affinis and the closely related Drosophila 150 athabasca), despite the addition of further mitochondrial and nuclear sequences to 151 increase confidence. In all cases we confirmed that viruses were only present in cDNA 152 and not in non reverse-transcription (RT) controls (i.e. DNA) by PCR, and so they cannot 153 be integrated into the insect genome (i.e. endogenous virus elements or EVEs [17]). COI 154 primers were used as a positive control for the presence of DNA in the non RT template. 155 bioRxiv preprint doi: https://doi.org/10.1101/020107; this version posted August 6, 2015. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. 5

156 We identified sigma virus sequences in RNA-seq data from Drosophila montana [22]. We 157 used RT-PCR on an infected fly line to amplify the virus sequence, and carried out 158 additional Sanger sequencing with primers designed using the RNA-seq assembly. 159 Additional virus sequences were identified from an RNA-seq analysis of pools of wild 160 caught Drosophila: DImmSV from Drosophila immigrans (collection and sequencing 161 described [23]), DTriSV from a pool of Drosophila tristis and SDefSV from 162 Scaptodrosophila deflexa (both Darren Obbard, unpublished data), accessions in tables 163 S1 and S2 (http://dx.doi.org/10.6084/m9.figshare.1502665). 164 165 Discovery of rhabdoviruses in public sequence databases 166 167 Rhabdovirus L gene sequences were used to search (tblastn) against expressed 168 sequence tag (EST) and transcriptome shotgun assembly (TSA) databases (NCBI). All 169 hits were reciprocally BLAST searched against Genbank cDNA and RefSeq databases and 170 only retained if they hit a virus-like sequence. We used two approaches to examine 171 whether sequences were present as RNA but not DNA. First, where assemblies of whole- 172 genome shotgun sequences were available, we used BLAST to test whether sequences 173 were integrated into the host genome. Second, for the virus sequences in the butterfly 174 Pararge aegeria and the medfly Ceratitis capitata we were able to obtain infected 175 samples to confirm the sequences are only present in RNA by performing PCR on both 176 genomic DNA and cDNA as described above (samples kindly provided by Casper 177 Breuker/Melanie Gibbs, and Philip Leftwich respectively) 178 179 Phylogenetic analysis 180 181 All available rhabdovirus-like sequences were downloaded from Genbank (accessions in 182 Data S2: http://dx.doi.org/10.6084/m9.figshare.1425419). Amino acid sequences for 183 the L gene (encoding the RNA Dependent RNA Polymerase or RDRP) were used to infer 184 the phylogeny (L gene sequences: http://dx.doi.org/10.6084/m9.figshare.1425067), as 185 they contain conserved domains that can be aligned across this diverse group of viruses. 186 Sequences were aligned with MAFFT [24] under default settings and then poorly aligned 187 and divergent sites were removed with either TrimAl (v1.3 strict settings, implemented 188 on Phylemon v2.0 server, alignment: http://dx.doi.org/10.6084/m9.figshare.1425069) 189 [25] or Gblocks (v0.91b selecting smaller final blocks, allowing gap positions and less 190 strict flanking positions to produce a less stringent selection, alignment: 191 http://dx.doi.org/10.6084/m9.figshare.1425068) [26]. These resulted in alignments of 192 1492 and 829 amino acids respectively. 193 194 Phylogenetic trees were inferred using Maximum Likelihood in PhyML (v3.0) [27] using 195 the LG substitution model [28] (preliminary analysis confirmed the results were robust 196 to the amino acid substitution model selected), with a gamma distribution of rate 197 variation with four categories and using a sub-tree pruning and regrafting topology 198 searching algorithm. Branch support was estimated using Approximate Likelihood-Ratio 199 Tests (aLRT) that is reported to outperform bootstrap methods [29]. Figures were 200 created using FIGTREE (v. 1.4) [30]. 201 202 Analysis of genetic structure between viruses taken from different hosts and ecologies bioRxiv preprint doi: https://doi.org/10.1101/020107; this version posted August 6, 2015. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. 6

203 204 We measured the degree of genetic structure between virus sequences identified in 205 different categories of host (arthropods, vertebrates and plants) and ecosystems 206 (terrestrial and aquatic). We measured the degree of genetic structure between virus

207 sequences from different groups of hosts/ecosystems using Hudson’s Fst estimator [31]

208 following the recommendations of [32]. We calculated Fst as: " % 209 number of differences within a population Fst =1−$ ' # number of differences between populations& 210 where a population is a host category or ecosystem. The significance of this value was 211 tested by permuting the host/ecosystem association of viruses. Since arthropods are the 212 most sampled host, we tested for evidence of genetic structure within the arthropod- 213 associated viruses that would suggest co-divergence with their hosts or preferential 214 host-switching between closely related hosts. We calculated the Pearson correlation 215 coefficient of the evolutionary distances between viruses and the evolutionary distances 216 between their hosts and tested for significance by permutation (as in [33]). We used the 217 patristic distances of our ML tree for the virus data and a time-tree of arthropod genera, 218 using published estimates of divergence dates [34, 35]. 219 220 Reconstruction of host associations 221 222 Viruses were categorised as having one of four types of host association: arthropod- 223 specific, vertebrate-specific, arthropod-vectored plant, or arthropod-vectored 224 vertebrate. However, the host association of some viruses are uncertain when they have 225 been isolated from vertebrates, biting-arthropods or plant-sap-feeding arthropods. Due 226 to limited sampling it was not clear whether viruses isolated from vertebrates were 227 vertebrate specific or arthropod-vectored vertebrate viruses; or whether viruses 228 isolated from biting-arthropods were arthropod specific viruses or arthropod-vectored 229 vertebrate viruses; or if viruses isolated from plant-sap-feeding arthropods were 230 arthropod-specific or arthropod-vectored plant viruses. 231 232 We omitted three viruses that were isolated from hosts outside of these four categories 233 (viruses from a nematode, fungus and cnidarian) from our analyses. We classified three 234 of the fish infecting dimarhabdoviruses as vertebrate specific based on the fact they can 235 be transmitted via immersion in water containing virus during experimental conditions 236 [36-38], and the widely held belief amongst the fisheries community that these viruses 237 are not typically vectored [8]. However, there is some evidence these viruses can be 238 transmitted by arthropods (sea lice) in experiments [8, 9] and so we would recommend 239 this be interpreted with some caution. Additionally, although we classified the viruses 240 identified in sea-lice as having biting arthropod hosts, they may be crustacean-specific. 241 The two viruses from Lepeophtheirus salmonis do not seem to infect the fish they 242 parasitise and are present in all developmental stages of the lice, suggesting they may be 243 transmitted vertically [39]. 244 245 We simultaneously estimated both the current and ancestral host associations, and the 246 phylogeny of the viruses, using a Bayesian analysis, implemented in BEAST v1.8 [40, 41]. 247 Since accurate branch lengths are essential for this analysis, we used a subset of the bioRxiv preprint doi: https://doi.org/10.1101/020107; this version posted August 6, 2015. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. 7

248 sites and strains used in the Maximum Likelihood (ML) analysis. We retained 189 taxa; 249 all rhabdoviruses excluding the divergent fish-infecting novirhabdovirus clade and the 250 virus from Hydra, as well as the viruses from Lolium perenne and Conwentzia 251 psociformis, which had a large number of missing sites. Sequences were trimmed to a 252 conserved region of 414 amino acids where data was recorded for most of these viruses 253 (the Gblocks alignment trimmed further by eye: 254 http://dx.doi.org/10.6084/m9.figshare.1425431). We used the host-association 255 categories described above, which included ambiguous states. To model amino acid 256 evolution we used an LG substitution model with gamma distributed rate variation 257 across sites [28] and an uncorrelated lognormal relaxed clock model of rate variation 258 among lineages [42]. To model the evolution of the host associations we used an 259 asymmetric transition rate matrix (allowing transitions to and from a host association to 260 take place at different rates) and a strict clock model. We also examined how often these 261 viruses jumped between different classes of hosts using reconstructed counts of 262 biologically feasible changes of host association and their HPD confidence intervals (CI) 263 using Markov Jumps [43]. These included switches between arthropod-specific and both 264 arthropod-vectored vertebrate and arthropod-vectored plant states, and between 265 vertebrate specific and arthropod-vectored vertebrate states. We used a constant 266 population size coalescent prior for the relative node ages (using a birth-death prior 267 gave equivalent results) and the BEAUti v1.8 default priors for all other parameters [40] 268 (BEAUti xml http://dx.doi.org/10.6084/m9.figshare.1431922). Convergence was 269 assessed using Tracer v1.6 [44], and a burn-in of 30% was removed prior to the 270 construction of a consensus tree, which included a description of ancestral host 271 associations. High effective sample sizes were achieved for all parameters (>200). 272 273 The maximum clade credibility tree estimated 274 (http://dx.doi.org/10.6084/m9.figshare.1425436) for the host association 275 reconstruction was very similar to the independently estimated maximum likelihood 276 phylogeny (http://dx.doi.org/10.6084/m9.figshare.1425083), which made no 277 assumptions about the appropriateness or otherwise of applying a clock model. The 278 minor topological differences may be expected in reconstructions that differ in their 279 assumptions about evolutionary rates. In Figure 2 we have transferred the ancestral 280 state reconstruction from the BEAST tree to the maximum likelihood tree. 281 282 283 Results 284 285 Novel rhabdoviruses from RNA-seq 286 287 To search for new rhabdoviruses we collected a variety of different species of flies,

288 screened them for CO2 sensitivity, which is a common symptom of infection, and 289 sequenced total RNA of these flies by RNA-seq. We identified rhabdovirus-like 290 sequences from a de-novo assembly by BLAST, and used PCR to identify which samples 291 these sequences came from. 292 293 This approach resulted in eleven rhabdovirus-like sequences from nine (possibly ten) 294 species of fly. Seven of these viruses were previously unknown and four had been bioRxiv preprint doi: https://doi.org/10.1101/020107; this version posted August 6, 2015. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. 8

295 reported previously from shorter sequences (Tables S1 and S2 296 http://dx.doi.org/10.6084/m9.figshare.1502665). The novel viruses were highly 297 divergent from known viruses. Sigma viruses known from other species of Drosophila 298 typically have genomes of ~12.5Kb [12, 45], and six of our sequences were 299 approximately this size, suggesting they are near-complete genomes. None of the 300 viruses discovered in our RNA-seq data were integrated into the host genome (see 301 Methods for details). 302 303 To investigate the putative gene content of the viruses, we predicted putative genes 304 based on open reading frames (ORFs). For the viruses with apparently complete 305 genomes (Figure 1), we found that those from Drosophila ananassae, Drosophila affinis, 306 Drosophila immigrans and Drosophila sturtvanti contained ORFs corresponding to the 307 five core genes found across all rhabdoviruses, with an additional ORF between the P 308 and M genes. This is the location of the X gene found in sigma viruses, and in three of the 309 four novel viruses it showed BLAST sequence similarity to the X gene of sigma viruses. 310 The virus from Drosophila busckii did not contain an additional ORF between the P and 311 M genes, but instead contained an ORF between the G and L gene. The high mutation 312 rate of RNA viruses mean that such ORFs are highly likely to be functional and retained 313 by selection, as non-function ORFs would quickly be degraded. 314 315 Using the phylogeny described below, we have classified our newly discovered viruses 316 as either sigma viruses, rhabdoviruses or other viruses, and named them after the host 317 species they were identified from (Figure 1) [46]. We also found one other novel 318 mononegavirales-like sequence from Drosophila unispina that groups with a recently 319 discovered clade of arthropod associated viruses (Nyamivirus clade [13], see Table S5 320 and the full phylogeny: http://dx.doi.org/10.6084/m9.figshare.1425083), as well as five 321 other RNA viruses from various families (data not shown), confirming our approach can 322 detect a wide range of divergent viruses. 323

324 325 Figure 1. Genome organization of newly discovered viruses. 326 Putative genes are shown in colour, non-coding regions are shown in black. ORFs were 327 designated as the first start codon following the transcription termination sequence (7 U’s) of the bioRxiv preprint doi: https://doi.org/10.1101/020107; this version posted August 6, 2015. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. 9

328 previous ORF to the first stop codon. Dotted lines represent parts of the genome not sequenced. 329 These viruses were either from our own RNA-seq data, or were first found in in public databases 330 and key features verified by PCR and Sanger sequencing. Rhabdovirus genomes are typically 331 ~11-13kb long and contain five core genes 3’-N-P-M-G-L-5’ [3]. However, a number of groups of 332 rhabdoviruses contain additional accessory genes and can be up to ~16kb long [14, 47]. 333 334 New rhabdoviruses from public databases 335 336 We identified a further 26 novel rhabdovirus-like sequences by searching public 337 databases of assembled RNA-seq data with BLAST. These included 19 viruses from 338 arthropods (Fleas, Crustacea, Lepidoptera, Diptera), one from a Cnidarian (Hydra) and 5 339 from plants (Table S3). Of these viruses, 19 had sufficient amounts of coding sequence 340 (>1000bp) to include in the phylogenetic analysis (Table S3), whilst the remainder were 341 too short (Table S4). 342 343 Four viruses from databases had near-complete genomes based on their size. These 344 were from the moth Triodia sylvina, the house fly Musca domestica (99% nucleotide 345 identity to Wuhan house fly virus 2 [13]), the butterfly Pararge aegeria and the medfly 346 Ceratitis capitata, all of which contain ORFs corresponding to the five core rhabdovirus 347 genes. The sequence from C. capitata had an additional ORF between the P and M genes 348 with BLAST sequence similarity to the X gene in sigma viruses. There were several 349 unusual sequences. Firstly, in the virus from P. aegeria there appear to be two full- 350 length glycoprotein ORFs between the M and L genes (we confirmed by Sanger 351 sequencing that both exist and the stop codon between the two genes was not an error). 352 Secondly, the Agave tequilana transcriptome contained a L gene ORF on a contig that 353 was the length of a typical rhabdovirus genome but did not appear to contain typical 354 gene content, suggesting it has very atypical genome organization, or has been 355 misassembled, or is integrated into its host plant genome [48]. Finally, the virus from 356 Hydra magnipapillata contained six predicted genes, but the L gene (RDRP) ORF was 357 unusually long. Some of the viruses we detected may be EVEs inserted into the host 358 genome and subsequently expressed [18]. For example, this is likely the case for the 359 sequence from the silkworm Bombyx mori that we also found in the silkworm genome, 360 and the L gene sequence from Spodoptera exigua that contains stop codons. Under the 361 assumption that viruses integrated into host genomes once infected those hosts, this 362 does not affect our conclusions below about the host range of these viruses [15-17]. We 363 also found nine other novel mononegavirale-like sequences that group with recently 364 discovered clades of insect viruses [13] (see Table S5 and 365 http://dx.doi.org/10.6084/m9.figshare.1425083). 366 367 Rhabdovirus Phylogeny 368 369 To reconstruct the evolution of the Rhabdoviridae we have produced the most complete 370 phylogeny of the group to date (Figure 2). We aligned the relatively conserved L gene 371 (RNA Dependant RNA Polymerase) from our newly discovered viruses with sequences 372 of known rhabdoviruses to give an alignment of 195 rhabdoviruses (and 26 other 373 mononegavirales as an outgroup). We reconstructed the phylogeny using different 374 sequence alignments and methodologies, and these all gave qualitatively similar results bioRxiv preprint doi: https://doi.org/10.1101/020107; this version posted August 6, 2015. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. 10

375 with the same major clades being reconstructed (Gblocks: 376 http://dx.doi.org/10.6084/m9.figshare.1425083, TrimAl: 377 http://dx.doi.org/10.6084/m9.figshare.1425082 and BEAST: 378 http://dx.doi.org/10.6084/m9.figshare.1425436). The ML and Bayesian relaxed clock 379 phylogenies were very similar, suggesting our analysis is robust to the assumptions of a 380 relaxed molecular clock. The branching order between the clades in the 381 dimarhabdovirus supergroup was generally poorly supported and differed between the 382 methods and alignments. Eight sequences that we discovered were not included in this 383 analysis as they were considered too short, but their closest BLAST hits are listed in 384 Table S4 (http://dx.doi.org/10.6084/m9.figshare.1502665). 385 386 We recovered all of the major clades described previously (Figure 2), and found that the 387 majority of known rhabdoviruses belong to the dimarhabdovirus clade (Figure 2b). The 388 RNA-seq viruses from Drosophila fall into either the sigma virus clade (Figure 2b) or the 389 arthropod clade sister to the cyto- and nucleo- rhabdoviruses (Figure 2a). The viruses 390 from sequence databases are diverse, coming from almost all of the major clades with 391 the exception of the lyssaviruses. 392 393 Predicted host associations of viruses 394 395 With a few exceptions, rhabdoviruses are either arthropod-vectored viruses of plants or 396 vertebrates, or are vertebrate- or arthropod- specific. In many cases the only 397 information about a virus is the host from which it was isolated. Therefore, a priori, it is 398 not clear whether viruses isolated from vertebrates are vertebrate-specific or 399 arthropod-vectored, or whether viruses isolated from biting arthropods (e.g. 400 mosquitoes, sandflies, ticks, midges and sea lice) are arthropod specific or also infect 401 vertebrates. Likewise, it is not clear whether viruses isolated from sap-sucking insects 402 (all Hemiptera: aphids, leafhoppers, scale insect and mealybugs) are arthropod-specific 403 or arthropod-vectored plant viruses. 404 405 By combining data on the ambiguous and known host associations with phylogenetic 406 information, we are able to predict both the ancestral and present host associations of 407 viruses (http://dx.doi.org/10.6084/m9.figshare.1425436). To do this we used a 408 Bayesian phylogenetic analysis that simultaneously estimated the phylogeny and host 409 association of our data. In the analysis we defined our host associations either as 410 vertebrate-specific, arthropod-specific, arthropod-vectored vertebrate and arthropod- 411 vectored plant, or as ambiguous between two of these four states (see Methods). 412 413 This approach identified a large number of viruses that are likely to be new arthropod- 414 vectored vertebrate viruses (Figure 2b). 80 of 89 viruses with ambiguous host 415 associations were assigned a host association with strong posterior support (>0.95). Of 416 the 52 viruses found in biting arthropods, 45 were predicted to be arthropod-vectored 417 vertebrate viruses, and 6 to be arthropod-specific. Of the 30 viruses found in 418 vertebrates, 22 were predicted to be arthropod-vectored vertebrate viruses, and 2 were 419 predicted to be vertebrate-specific (both fish viruses). Of the 7 viruses found in plant- 420 sap-feeding arthropods (Figure 2a), 3 were predicted to be plant-associated and 2 421 arthropod-associated. bioRxiv preprint doi: https://doi.org/10.1101/020107; this version posted August 6, 2015. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. 11

Drosophila unispina virus DrosophilaDrosophila unispina unispina virus virus Xincheng Mosquito Virus XinchengXincheng Mosquito Mosquito Virus Virus ? ! Shuangao Fly Virus 2 ShuangaoShuangao Fly Virus Fly 2 Virus 2 Corydalus cornutus TSA CorydalusCorydalus cornutus cornutus TSA TSA ? ! Ganaspis sp G1 TSA GanaspisGanaspis sp G1 TSA sp G1 TSA Stylops melittae TSA Stylops Stylopsmelittae melittaeTSA TSA ? ! Tacheng Tick Virus 6 TachengTacheng Tick Virus Tick 6 Virus 6 A ? ! Sanxia Water Strider Virus 4 Sanxia WaterSanxia Strider Water Virus Strider 4 Virus 4 Meligethes aeneus TSA MeligethesMeligethes aeneus aeneusTSA TSA ? ! Lishi Spider Virus 2 Lishi SpiderLishi Virus Spider 2 Virus 2 Trialeurodes vaporariorum TSA TrialeurodesTrialeurodes vaporariorum vaporariorum TSA TSA Bemisia tabaci TSA BemisiaBemisia tabaci TSA tabaci TSA

Hydra magnipapillata TSA Hydra (Cnidarian) ? ! Viral hemorrhagic septicemia virus Fish Snakehead rhabdovirus Fish Infectious haematopoietic necrosis virus Fish novirhabdoviruses Hirame rhabdovirus Fish Soybean cyst nematode virus Nematode Shuangao Insect Virus 6 A Diptera and Lepidoptera Taastrup virus AP Leafhopper ! Drosophila sturtvanti rhabdovirus A Drosophilid fruit fly ? ! Drosophila subobscura rhabdovirus A Drosophilid fruit fly Wuhan House Fly Virus 2 A Muscid house fly ? ! Musca domestica TSA A Muscid house fly Shayang Fly Virus 3 A Diptera species (Calliphorid and Muscid flies) ? ! Wuhan Fly Virus 3 A Diptera species (Calliphorid and Sacophagid flies) Frankliniella occidentalis TSA A Western flower thrip Wuhan Ant Virus A Japanese carpenter ant Triodia sylvina TSA A Orange swift moth Spodoptera frugiperda rhabdovirus A Fall army worm moth Wuhan Mosquito Virus 9 BA Mosquitoes Shuangao Bedbug Virus 2 BA Bedbug Oropsylla silantiewi TSA BA Flea Sanxia Water Strider Virus 5 A Water Strider Drosophila busckii rhabdovirus A Drosophilid fruit fly Jingshan Fly Virus 2 A Sarcophagid flesh fly Tacheng Tick Virus 7 BA Ticks Farmington virus V Bird species Kerria lacca TSA AP Scale insect Planococcus citri TSA AP Citrus mealybug Northern cereal mosaic virus P Cereals and leafhoppers ? ! Wuhan Insect virus 4 AP Aphid or its parasitoid wasp Lettuce yellow mottle virus P Lettuce and aphid Lettuce necrotic yellows virus P Lettuce, other dicot plants and aphids cyto- and nucleo-

Wuhan Insect virus 6 AP Aphid or its parasitoid wasp rhabdoviruses Medicago sativa TSA P Alfalfa ? ! Wuhan Insect virus 5 AP Aphid or its parasitoid wasp Persimmon virus A P Persimmon tree Lolium perenne TSA P Rye grass Maize fine streak virus P Maize and leafhopper ? ? ! ! Orchid fleck virus P Flowering plant ? ! Lotus corniculatus TSA P Flowering plant Sonchus yellow net P Flowering plant and aphid ? ! Eggplant mottled dwarf virus P Eggplant Rice yellow stunt virus P Rice and leafhoppers Agave tequilana TSA P Flowering plant Maize Iranian mosaic virus P Cereals and planthopper Taro vein chlorosis virus P Taro Maize mosaic virus P Maize and planthoppers Fox fecal rhabdovirus V Fox (fecal sample) Ikoma lyssavirus VS African Civets West Caucasian bat virus VS Bats Shimoni bat virus VS Bats Mokola virus 86101RCA VS Mammals species Mokola virus isolate 86100CAM VS Mammals species l Lagos bat virus KE131 VS Bats y

? ! s Lagos bat virus i8619NGA VS Mammals species s

Rabies virus VS Mammals species including humans a

Lyssavirus Ozernoe VS Humans v i Planococcus citri TSA Irkut virus A Citrus mealybug VS Bats r Northern cereal mosaic virusDuvenhage virus 86132SA P Cereals and leafhoppersVS Humans and bats u European bat lyssavirus 1 8918FRA VS Mammals species s

Wuhan Insect virus 4 A Aphid or its parasitoid wasp e ? ? European bat lyssavirus RV9 1 VS Bats

? Lettuce yellow mottle! virus P Lettuce and aphid s Lettuce necrotic yellows virusAravan virus P Lettuce, other dicotVS plantsBats and aphids Wuhan Insect virus 6 Bokeloh bat lyssavirus A Aphid or its parasitoidVS waspBats Medicago sativa TSA European bat lyssavirus 2 9018HOLP Alfalfa VS Humans and bats ? ? Khujand lyssavirus VS Bats ? Wuhan Insect virus! 5 A Aphid or its parasitoid wasp Persimmon virus A Australian bat lyssavirus a P Persimmon tree VS Bats and humans Lolium perenne TSA Australian bat lyssavirus b P Rye grass VS Bats and humans Maize fine streak virus Arboretum virus P Maize and leafhopperBA Mosquitoes Puerto Almendras virus BA Mosquitoes ? ? ? ? Orchid fleck virus P Flowering plant 0.4 ? ? Lotus corniculatus TSAMuir Springs virus P Flowering plant BA Mosquitoes Sonchus yellow net Bahia Grande virus P Flowering plant andBA aphidMosquitoes Fig 2B HarlingenAssociated virus hostsBA Mosquitoes ? ? Eggplant mottled dwarf virus P Eggplant Rice yellow stunt virus Moussa virus P Rice and leafhoppersBA Mosquitoes Agave tequilana TSA New Minto virus P Flowering plant BA Ticks Maize Iranian mosaic virusSawgrassArthropod-vectored virus P Cereals and planthopperBA Ticksplant Taro vein chlorosis virus Connecticut virus P Taro VV Ticks and rabbits Maize mosaic virus Long Island tick rhabdovirus P Maize and planthoppersBA Ticks Fox fecal rhabdovirus Taishun Tick Virus V Fox (fecal sample)BA Ticks HuangpiArthropods Tick Virus 3 BA Ticks ? Ikoma !lyssavirus VS African Civets West Caucasian bat virusBole Tick Virus 2 VS Bats BA Ticks Shimoni bat virus Tacheng Tick Virus 3 VS Bats BA Ticks Mokola virus 86101RCA WuhanV Tickertebrate Virus 1 VS Mammalsspecific speciesBA Ticks Mokola virus isolate 86100CAMHumulus lupulus TSA VS Mammals speciesP Hopes Lagos bat virus KE131 Beaumont virus VS Bats BA Mosquitoes ? ? Lagos bat virus i8619NGACulex tritaeniorhynchusArthropod-vectored rhabdovirusVS Mammals speciesBA Mosquitoesvertebrate Rabies virus Conwentzia psociformis TSAVS Mammals species includingLacewing humans 422 Lyssavirus Ozernoe North Creek Virus VS Humans BA Mosquitoes Durham virus V Birds ? !Irkut virus VS Bats Duvenhage virus 86132SAKlamath virus VS Humans and batsV Voles European bat lyssavirus 1Tupaia 8918FRA virus VS Mammals speciesV Tree shrews Kwatta virus VV Mosquitoes, ticks and mammals ? ? European bat lyssavirus RV9 1 VS Bats Aravan virus Oak Vale virus VS Bats VV Mosquitoes and swine Bokeloh bat lyssavirus Garba virus VS Bats V Birds B Sunguru virus V Domestic chickens European bat lyssavirus 2 9018HOL VS Humans and bats Fig 2A Khujand lyssavirus Almpiwar virus VS Bats V Lizards ? ? Australian bat lyssavirus aSripur virus VS Bats and humansBA Sandflies Niakha virus BA Sandflies ? Australian bat lyssavirus b ! VS Bats and humans ? Arboretum! virus Chaco virus BA Mosquitoes V Lizards Puerto Almendras virus Sena Madureira virus BA Mosquitoes V Lizards Muir Springs virus Caligus rogercresseyi 11114047BA TSAMosquitoes BA Sea louse Bahia Grande virus Caligus rogercresseyi 11125273BA TSAMosquitoes BA Sea louse Harlingen virus Lepeophtheirus salmonis rhabdovirusBA Mosquitoes 127 BA Sea louse Moussa virus Lepeophtheirus salmonis rhabdovirusBA Mosquitoes 9 BA Sea louse New Minto virus La Joya virus BA Ticks VV Mosquitoes and rodents Sawgrass virus Wongabel virus BA Ticks VV Midges and birds Connecticut virus Ord River virus VV Ticks and rabbitsBA Mosquitoes Long Island tick rhabdovirusParry Creek virus BA Ticks BA Mosquitoes Ngaingan virus VV Midges, cattle and macropods ? Taishun Tick Virus ! BA Ticks Huangpi Tick Virus 3 Marco virus BA Ticks V Lizards ? ? ? Bole Tick Virus 2 ! Joinjakaka virus BA Ticks VV Mosquitoes and cattle ? Tacheng Tick Virus 3 ! Gray Lodge virus BA Ticks BA Mosquitoes Wuhan Tick Virus 1 Landjia virus BA Ticks V Birds Humulus lupulus TSA Manitoba virus P Hops BA Mosquitoes Beaumont virus Mosqueiro virus BA Mosquitoes BA Mosquitoes Culex tritaeniorhynchus rhabdovirusHart Park virus BA Mosquitoes VV Mosquitoes and birds Conwentzia psociformisFlanders TSA virus Lacewing VV Mosquitoes and birds North Creek Virus Kamese virus BA Mosquitoes VV Mosquitoes and humans Durham virus Mossuril virus V Birds VV Mosquitoes, birds and mammals including humans ? ? Klamath virus Itacaiunas virus V Voles BA Midges Tupaia virus Iriri virus V Tree shrews BA Sandflies Curionopolis virus VV Midges and mammals ? Kwatta virus ! VV Mosquitoes, ticks and mammals Oak Vale virus Rochambeau virus VV Mosquitoes and swineBA Mosquitoes Garba virus Santa barbara virus V Birds A Psychodidae drain fly Sunguru virus Inhangapi virus V Domestic chickensVV Sandflies and rodents Almpiwar virus Aruac virus V Lizards VV Mosquitoes and birds Sripur virus Xiburema virus BA Sandflies BA Mosquitoes Niakha virus Bas Congo virus BA Sandflies V Humans ? ? Coastal Plains virus V Bovids ? ? Chaco virus V Lizards Sena Madureira virus Sweetwater Branch virus V Lizards VV Midges and cattle Caligus rogercresseyi 11114047Bivens Arm TSA virus BA Sea louse VV Midges and cattle ? Caligus rogercresseyi! 11125273Tibrogargan TSA virus BA Sea louse VV Midges and bovids Lepeophtheirus salmonis Koolpinyahrhabdovirus virus 127 BA Sea louse V Cattle Yata virus BA Mosquitoes ? Lepeophtheirus salmonis! rhabdovirus 9 BA Sea louse La Joya virus Adelaide River virus VV Mosquitoes and rodentsV Cattle Wongabel virus Kimberley virus VV Midges and birdsVV Midges, mosquitoes and cattle Ord River virus Malakal virus BA Mosquitoes BA Mosquitoes Parry Creek virus Berrimah virus BA Mosquitoes BA Cattle Bovine ephmeral fever virus VV Midges, mosquitoes and ruminants ? ? Ngaingan virus VV Midges, cattle and macropods Marco virus Drosophila ananassae sigmaV virusLizards A Drosophilid fruit fly Joinjakaka virus Pararge aegeria rhabdovirusVV Mosquitoes and cattleA Speckled wood butterfly ? ? Spodoptera exigua TSA A Beet army worm moth ? ? Gray Lodge virus BA Mosquitoes Landjia virus Wuhan Insect virus 7 V Birds A Aphid or its parasitoid wasp Manitoba virus Drosophila immigrans sigmaBA virus Mosquitoes A Drosophilid fruit fly Mosqueiro virus Drosophila obscura sigma virusBA Mosquitoes A Drosophilid fruit fly Hart Park virus Ceratitis capitata sigma virusVV Mosquitoes and birdsA Tephritid fruit fly ? ! Flanders virus Wuhan Louse Fly Virus 10 VV Mosquitoes and birdsBA Louse fly Kamese virus Wuhan Louse Fly Virus 8 VV Mosquitoes and humansBA Louse fly Mossuril virus Wuhan Louse Fly Virus 9 VV Mosquitoes, birdsBA and mammalsLouse fly including humans Itacaiunas virus Scaptodrosophila deflexa sigmaBA virusMidges A Drosophilid fruit fly ? Drosophila! melanogaster sigma virus HAP23 isolate A Drosophilid fruit fly

Iriri virus BA Sandflies dimarhabdovirus supergroup Curionopolis virus Drosophila melanogaster sigmaVV virus MidgesAP30 isolate and mammals A Drosophilid fruit fly ? ? Rochambeau virus Wuhan House Fly Virus 1 BA Mosquitoes A Muscid house fly Santa barbara virus Muscina stabulans sigma virusA Psychodidae drainA fly False stable fly ? ! Shayang Fly Virus 2 A Diptera species (Muscid house fly and Calliphorid laterine fly) ? Inhangapi virus ! VV Sandflies and rodents Aruac virus Wuhan Fly Virus 2 VV Mosquitoes and birdsA Muscid house fly Xiburema virus Drosophila sturtvanti sigmaBA virus Mosquitoes A Drosophilid fruit fly Bas Congo virus Drosophila tristis sigma virusV Humans A Drosophilid fruit fly Drosophila montana sigma virus A Drosophilid fruit fly ? Coastal Plains virus ! V Bovids Sweetwater Branch virusDrosophila affinis sigma virusVV Midges and cattleA Drosophilid fruit fly Bivens Arm virus Drosophila affinis or athabascaVV sigmaMidges virus and cattleA Drosophilid fruit fly Yongjia Tick Virus 2 BA Ticks ? ? ? Tibrogargan virus ! VV Midges and bovids Koolpinyah virus Nkolbisson virus V Cattle VV Mosquitoes and humans Yata virus Barur virus BA Mosquitoes VV Ticks, mosquitoes, fleas and mammals ? ? Adelaide River virus Nishimuro virus V Cattle V Wild boar Kimberley virus Fukuoka virus VV Midges, mosquitoesVV and cattleMidges, mosquitoes and cattle Malakal virus Kern Canyon virus BA Mosquitoes V Bats Berrimah virus Keuraliba virus BA Cattle V Rodents Bovine ephmeral fever virusOita virus VV Midges, mosquitoesV and ruminantsBats Drosophila ananassae sigmaFikirini virus bat rhabdovirus A Drosophilid fruit flyV Bats Pararge aegeria rhabdovirusWuhan Louse Fly Virus 5 A Speckled wood butterflyBA Louse fly Spodoptera exigua TSAMount Elgon bat virus A Beet army worm mothV Bats Scophthalmus maximus rhabdovirus V Cultured turbot ? Wuhan Insect virus! 7 A Aphid or its parasitoid wasp Drosophila immigrans sigmaDolphin virus rhabdovirus A Drosophilid fruit flyV Dolphins and porpoise Eel Virus European X V European eel ? Drosophila obscura sigma! virus A Drosophilid fruit fly s

Starry flounder rhabdovirus V Starry flounder i

? Ceratitis capitata sigma !virus A Tephritid fruit fly g Siniperca chuatsi virus V Mandarin fish ? ? Wuhan Louse Fly Virus 10 BA Louse fly m Wuhan Louse Fly Virus 8Spring viremia of carp virus BA Louse fly V Common carp Wuhan Louse Fly Virus 9Grass carp rhabdovirus BA Louse fly V Grass carp a

Scaptodrosophila deflexaPike sigma fry rhabdovirus virus A Drosophilid fruit flyV Northern pike v ? ? Tench rhabdovirus V Tench i Drosophila melanogaster sigma virus HAP23 isolate A Drosophilid fruit fly r Drosophila melanogaster Wuhansigma virus Louse AP30 Fly Virusisolate 11 A Drosophilid fruit flyBA Louse fly u Wuhan House Fly Virus 1Radi virus A Muscid house fly BA Sandflies s Yug Bogdanovac virus BA Sandflies e

Muscina stabulans sigma virus A False stable fly s ? ? Shayang Fly Virus 2 Jurona virus A Diptera species (MuscidBA houseMosquitoes fly and Calliphorid laterine fly) ? ? Wuhan Fly Virus 2 Malpais Spring virus A Muscid house fly BA Mosquitoes Chandipura virus VV Sandflies and mammals including humans ? Drosophila sturtvanti sigma !virus A Drosophilid fruit fly Drosophila tristis sigmaIsfahan virus virus A Drosophilid fruit flyVV Mosquitoes, ticks, sandflies, mammals including humans Drosophila montana sigmaPerinet virus virus A Drosophilid fruit flyBA Mosquitoes and sandflies ? Drosophila affinis? sigmaCarajas virus oncolytic virus A Drosophilid fruit flyBA Sandflies Drosophila affinis or athabasca sigma virus A Drosophilid fruit fly Yongjia Tick Virus 2 BA Ticks ? ? Nkolbisson virus VV Mosquitoes and humans Barur virus VV Ticks, mosquitoes, fleas and mammals Nishimuro virus V Wild boar Fukuoka virus VV Midges, mosquitoes and cattle Kern Canyon virus V Bats Keuraliba virus V Rodents Oita virus V Bats Fikirini bat rhabdovirus V Bats Wuhan Louse Fly Virus 5 BA Louse fly Mount Elgon bat virus V Bats Scophthalmus maximus rhabdovirus V Cultured turbot ? ? Dolphin rhabdovirus V Dolphins and porpoise Eel Virus European X V European eel ? ? ? Starry? flounder rhabdovirus V Starry flounder Siniperca chuatsi virus V Mandarin fish Spring viremia of carp virus V Common carp Associated hosts Grass carp rhabdovirus V Grass carp Pike fry rhabdovirus V Northern pike Tench rhabdovirus V Tench Plants Wuhan Louse Fly Virus 11 BA Louse fly Radi virus BA Sandflies Yug Bogdanovac virus BA Sandflies Arthropods Jurona virus BA Mosquitoes Malpais Spring virus BA Mosquitoes Chandipura virus VV Sandflies and mammals including humans ? Isfahan? virus VV Mosquitoes, ticks, sandflies, mammals including humans Vertebrate specific Perinet virus BA Mosquitoes and sandflies Carajas oncolytic virus BA Sandflies Vesicular stomatitis virus New Jersey Hazelhurst VV Mammals including humans, biting and non-biting diptera Arthropod-vectors and vertebrates Vesicular stomatitis virus New Jersey VV Mammals including humans, biting and non-biting diptera Maraba virus BA Sandflies Vesicular stomatitis virus Indiana VV Mammals including humans, sandflies and mosquitoes ? Morreton? virus BA Sandflies 0.4 ? Vesicular? stomatitis virus Cocal VV Mites, mosquitoes and mammals Vesicular stomatitis virus Alagoas Indiana 3 V Mammals including humans 423 0.4 0.4 0.4 424 Figure 2. Maximum likelihood phylogeny of the Rhabdoviridae. bioRxiv preprint doi: https://doi.org/10.1101/020107; this version posted August 6, 2015. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. 12

425 (A) shows the basal fish-infecting novirhabdoviruses, an unassigned group of arthropod 426 associated viruses, the plant infecting cyto- and nucleo- rhabdoviruses, as well as the vertebrate 427 specific lyssaviruses. (B) shows the dimarhabdovirus supergroup, which is predominantly 428 composed of arthropod-vectored vertebrate viruses, along with the arthropod-specific sigma 429 virus clade. Branches are coloured based on the Bayesian host association reconstruction 430 analysis. Black represents taxa omitted from host-state reconstruction or associations with <0.95 431 support. The tree was inferred from L gene sequences using the Gblocks alignment. The columns 432 of text are the virus name, the host category used for reconstructions, and known hosts (from left 433 to right). Codes for the host categories are: vs= vertebrate-specific, vv= arthropod-vectored 434 vertebrate, a= arthropod specific, ba = biting-arthropod (ambiguous state), v = vertebrate 435 (ambiguous state) and ap=plant-sap-feeding-arthropod (ambiguous state). Names in bold and 436 underlined are viruses discovered in this study. The tree is rooted with the Chuvirus clade (root 437 collapsed) as identified as an outgroup in [13] but we note this gives the same result as midpoint 438 and the molecular clock rooting. Nodes labelled with question marks represent nodes with aLRT 439 (approximate likelihood ratio test) statistical support values less than 0.75. Scale bar shows 440 number of amino-acid substitutions per site. 441 442 To test the ability of these reconstructions to correctly predict host associations we 443 randomly selected 10 viruses with known host associations, and assigned them to an 444 ambiguous state. This analysis correctly returned the true host association for all 10 of 445 the randomly chosen viruses with strong posterior support (mean support 0.97, range 446 0.92-1.00). 447 448 Ancestral host associations and host-switches 449 450 Viral sequences from arthropods, vertebrates and plants form distinct clusters in the

451 phylogeny (Figure 2). To quantify this genetic structure we calculated a variant of the Fst 452 statistic between the sequences of viruses from different groups of hosts. There is 453 strong evidence of genetic differentiation between the sequences from different host 454 groups (P<0.001, Figure S1 http://dx.doi.org/10.6084/m9.figshare.1495351). 455 456 Our Bayesian analysis allowed us to infer the ancestral host association of 178 of 188 of 457 the internal nodes on the phylogenetic tree (support >0.95). A striking pattern that 458 emerged is that switches between major groups of hosts have occurred rarely during 459 the evolution of the rhabdoviruses (Figure 2). There are a few rare transitions on 460 terminal branches (Santa Barbara virus and the virus identified from the plant Humulus 461 lupulus) but these could represent errors in the host assignment (e.g. cross-species 462 contamination) as well as recent host shifts. Our analysis allows us to estimate the 463 number of times the viruses have switched between major host groups while accounting 464 for uncertainty about ancestral states, and the tree topology and root. There were only 465 two types of host-switch that we found evidence for. Our best estimate of the number of 466 switches from being an arthropod-vectored vertebrate virus to being arthropod specific 467 was three, although one is a jump on a terminal branch may represent contamination 468 (best modal estimate = 3, median = 3.1, CI’s= 1.9–6.1). Similarly, we estimate there have 469 been three transitions from being an arthropod-vectored vertebrate virus to a 470 vertebrate-specific virus (best modal estimate = 3, median = 3.1, CI’s= 2.9–5.3). We 471 could not determine the direction of the host shifts into the other host groups. 472 bioRxiv preprint doi: https://doi.org/10.1101/020107; this version posted August 6, 2015. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. 13

473 Vertebrate–specific viruses have arisen once in the lyssaviruses clade [3], as well as at 474 least once in fish dimarhabdoviruses (in one of the fish-infecting clades it is unclear if it 475 is vertebrate-specific or vector-borne from our reconstructions). There has also likely 476 been a single transition to being arthropod-vectored vertebrate viruses in the 477 dimarhabodovirus clade. 478 479 Insect-vectored plant viruses have arisen once in the cyto- and nucleo- rhabdoviruses, 480 although the ancestral state of these viruses is uncertain. A single virus identified from 481 the hop plant Humulus lupulus appears to fall within the dimarhabdovirus clade. 482 However, this may be because the plant was contaminated with insect matter, as the 483 same RNA-seq dataset contains COI sequences with high similarity to thrips. 484 485 There are two large clades of arthropod-specific viruses. The first is a sister group to the 486 large plant virus clade. This novel group of largely insect-associated viruses are 487 associated with a broad range of insects, including flies, butterflies, moths, ants, thrips, 488 bedbugs, fleas, mosquitoes, water striders and leafhoppers. The mode of transmission 489 and biology of these viruses is yet to be examined. The second clade of insect-associated 490 viruses is the sigma virus clade [11, 12, 19, 45]. These are derived from vector-borne 491 dimarhabdoviruses that have lost their vertebrate host and become vertically 492 transmitted viruses of insects [10]. They are common in Drosophilidae, and our results 493 suggest that they may be widespread throughout the Diptera, with occurrences in the 494 Tephritid fruit fly Ceratitis capitata, the stable fly Muscina stabulans, several divergent 495 viruses in the housefly Musca domestica and louse flies removed from the skin of bats. 496 For the first time we have found sigma-like viruses outside of the Diptera, with two 497 Lepidoptera associated viruses and a virus from an aphid/parasitoid wasp. All of the 498 sigma viruses characterised to date have been vertically transmitted [10], but some of 499 the recently described viruses may be transmitted horizontally – it has been speculated 500 that the viruses from louse flies may infect bats [49] and Shayang Fly Virus 2 has been 501 reported in two fly species [13] (although contamination could also explain this 502 result).Drosophila sigma virus genomes are characterised by an additional X gene 503 between the P and M genes [45]. Interestingly the two louse fly viruses with complete 504 genomes, Wuhan insect virus 7 from an aphid/parasitoid and Pararge aegeria 505 rhabdovirus do not have a putative X gene. The first sigma virus was discovered in 506 Drosophila melanogaster in 1937 [50], in the last few years related sigma viruses have 507 been found in other Drosophila species and a Muscid fly [10-12, 45]. Here we have found 508 sigma-like viruses in a diverse array of Diptera species, as well as other insect orders. 509 Overall, our results suggest sigma-like viruses may be associated with a wide array of 510 insect species. 511 512 Within the arthropod-associated viruses (the most sampled host group) it is common to 513 find closely related viruses in closely related hosts (Figure 1). This is reflected in a 514 positive correlation between the evolutionary distance between the viruses and the 515 evolutionary distance between their arthropod hosts (Pearson’s correlation=0.36, 95% 516 CI’s=0.34-0.38, P<0.001, Figure 3 and Figure S2 517 http://dx.doi.org/10.6084/m9.figshare.1495351). Since the virus phylogeny is 518 incongruent with that of the respective hosts, this suggests rhabdoviruses preferentially 519 host shift between closely related species [51, 52]. 1000

bioRxiv preprint doi: https://doi.org/10.1101/020107; this version posted August 6, 2015. The copyright holder for this preprint (which was1000 not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. 14 800

10001000# 800 # # # #

800 800# 600 CountCount # # # 600 600600# Frequency Count # #

# 400 400400# 7.5# # 1200# # 400 # 200200# # # 200 # 00# 0 200 520 521 Figure 3. The relationship between the evolutionary distance between viruses and the 522 evolutionary distance between their arthropod hosts (categorised by genus). Closely 523 related viruses tend to be found in closely related hosts. Permutation tests find a significant 0 524 positive correlation (correlation=0.36, 95% CI’s=0.34-0.38, P<0.001) between host and virus 525 evolutionary distance (see Figure S2). 526 0 527 We also find viruses clustering on the phylogeny based on the ecosystem of their hosts; 528 there is strong evidence of genetic differentiation between viruses from terrestrial and 529 aquatic hosts (P=0.007, Figure S3 http://dx.doi.org/10.6084/m9.figshare.1495351). 530 There has been one shift from terrestrial to aquatic hosts during the evolution of the 531 basal novirhabdoviruses, which have a wide host range in fish. There have been other 532 terrestrial to aquatic shifts in the dimarhabdoviruses: in the clades of fish and cetacean 533 viruses and the clade of viruses from sea-lice. 534 535 536 Discussion 537 538 Viruses are ubiquitous in nature and recent developments in high-throughput 539 sequencing technology have led to the discovery and sequencing of a large number of 540 novel viruses in arthropods [13, 14]. Here we have identified 43 novel virus-like 541 sequences, from our own RNA-seq data and public sequence repositories. Of these, 32 542 were rhabdoviruses, and 26 of these were from arthropods. Using these sequences we 543 have produced the most extensive phylogeny of the Rhabdoviridae to date, including a 544 total of 195 virus sequences. 545 546 In most cases we know nothing about the biology of the viruses beyond the host they 547 were isolated from, but our analysis provides a powerful way to predict which are 548 vector-borne viruses and which are specific to vertebrates or arthropods. We have bioRxiv preprint doi: https://doi.org/10.1101/020107; this version posted August 6, 2015. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. 15

549 identified a large number of new likely vector-borne viruses – of 85 rhabdoviruses 550 identified from vertebrates or biting insects we predict that 76 are arthropod-borne 551 viruses of vertebrates (arboviruses). By assigning viruses with known associations to an 552 ambiguous state, we have demonstrated that this analysis is robust and can correctly 553 predict the host associations of viruses. Along with the known arboviruses, this suggests 554 the majority of known rhabdoviruses are arboviruses, and all of these fall in a single 555 clade known as the dimarhabdoviruses. In addition to the arboviruses, we also 556 identified two clades of likely insect-specific viruses associated with a wide range of 557 species, suggesting rhabdoviruses may be associated with a wide array of arthropods 558 559 We found that shifts between distantly related hosts are rare in the rhabdoviruses, 560 which is perhaps unsurprising as both rhabdoviruses of vertebrates (rabies virus in 561 bats) and invertebrates (sigma viruses in Drosophildae) show a declining ability to 562 infect hosts more distantly related to their natural host [52-54]. It is thought that sigma 563 viruses may sometimes jump into distantly related but highly susceptible species [51, 564 52, 55], but our results suggest that this rarely happens between major groups such as 565 vertebrates and arthropods. It is nonetheless surprising that arthropod-specific viruses 566 have arisen rarely, as one might naively assume that there would be fewer constraints 567 on vector-borne viruses losing one of their hosts. Within the major clades, closely 568 related viruses often infect closely related hosts (Figure 2). For example, within the 569 dimarhabdoviruses viruses identified from mosquitoes, ticks, Drosophila, Muscid flies, 570 Lepidoptera and sea-lice all tend to cluster together (Figure 2B). However, it is also 571 clear that the virus phylogeny does not mirror the host phylogeny, and our data on the 572 clustering of hosts across the virus phylogeny suggest that following major transitions 573 between distantly related host taxa, viruses preferentially shift between more closely 574 related species (Figures 3, S1 and S2) in the same environment (Figure S3). 575 576 There has been a near four-fold increase in the number of recorded rhabdovirus 577 sequences in the last five years. In part this may be due to the falling cost of sequencing 578 transcriptomes [56], and initiatives to sequence large numbers of insect and other 579 arthropods [35]. The use of high-throughput sequencing technologies should reduce the 580 likelihood of sampling biases associated with PCR based discovery where people look 581 for similar viruses in related hosts. This suggests that the pattern of viruses forming 582 clades based on the host taxa they infect is likely to be robust. However, these efforts are 583 disproportionately targeted at arthropods, and it is possible that there may be a great 584 undiscovered diversity of viruses in other organisms [57]. 585 586 Rhabdoviruses infect a diverse range of host species, including a large number of 587 arthropods. Our search has unearthed a large number of novel rhabdovirus genomes, 588 suggesting that we are only just beginning to uncover the diversity of these viruses. The 589 host use by these viruses has been highly conserved during their evolution, which 590 provides a powerful tool to identify previously unknown arboviruses. Given the large 591 number of viruses currently being discovered through metagenomic studies [13, 58, 59], 592 in the future we will be faced by an increasingly large number of viral sequences with 593 little knowledge of the biology of the virus. Our phylogenetic approach can be extended 594 to predict key biological traits in other groups of novel pathogens where our knowledge 595 is incomplete. However, the rapid evolution of RNA viruses may mean traits may change bioRxiv preprint doi: https://doi.org/10.1101/020107; this version posted August 6, 2015. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. 16

596 over short timescales, and such an approach should complement, and not replace, 597 examining the basic biology of novel viruses. 598 599 Acknowledgments 600 601 Many thanks to Mike Ritchie for providing the DMonSV infected fly line; Casper Breuker 602 and Melanie Gibbs for PAegRV samples and Philip Leftwich for CCapSV samples. Thanks 603 to everyone who provided fly collections. 604 605 Contributions 606 607 BL and FMJ conceived and designed the study. BL and JD carried out molecular work. BL, 608 WJP, DJP and DJO carried out bioinformatic analysis. BL, GGRM and JJW carried out 609 phylogenetic analysis. BL and FMJ wrote the manuscript with comments from all other 610 authors. All authors gave final approval for publication. 611 612 Funding 613 614 BL and FMJ are supported by a NERC grant (NE/L004232/1), a European Research 615 Council grant (281668, DrosophilaInfection), a Junior Research Fellowship from Christ’s 616 College, Cambridge (BL). GGRM is supported by an MRC studentship. The metagenomic 617 sequencing of viruses from D. immigrans, D. tristis and S. deflexa was supported by a 618 Wellcome Trust fellowship (WT085064) to DJO. 619 620 621 References 622 623 1. Lipkin W.I., Anthony S.J. 2015 Virus hunting. Virology 479-480C, 194-199. 624 (doi:10.1016/j.virol.2015.02.006). 625 2. Liu S., Vijayendran D., Bonning B.C. 2011 Next generation sequencing 626 technologies for insect virus discovery. Viruses 3(10), 1849-1869. 627 (doi:10.3390/v3101849). 628 3. Dietzgen R.G., Kuzmin I.V. 2012 Rhabdoviruses: Molecular Taxonomy, Evolution, 629 Genomics, Ecology, Host-Vector Interactions, Cytopathology and Control. Norfolk, UK, 630 Caister Academic Press. 631 4. Bourhy H., Cowley J.A., Larrous F., Holmes E.C., Walker P.J. 2005 Phylogenetic 632 relationships among rhabdoviruses inferred using the L polymerase gene. Journal of 633 General Virology 86, 2849-2858. 634 5. Hampson K., Coudeville L., Lembo T., Sambo M., Kieffer A., Attlan M., Barrat J., 635 Blanton J.D., Briggs D.J., Cleaveland S., et al. 2015 Estimating the global burden of 636 endemic canine rabies. PLoS Negl Trop Dis 9(4), e0003709. 637 (doi:10.1371/journal.pntd.0003709). 638 6. Walker P.J., Blasdell K.R., Joubert D.A. 2012 Ephemeroviruses: Athropod-borne 639 Rhabdoviruses of Ruminants, with Large, Complex Genomes. In Rhabdoviruses: 640 Molecular Taxonomy, Evolution, Genomics, Ecology, Host-Vector Interactions, 641 Cytopathology and Control (eds. Dietzgen R.G., Kuzmin I.V.), pp. 59-88. Norfolk, UK, 642 Caister Academic Press. 643 7. Hogenhout S.A., Redinbaugh M.G., Ammar E.D. 2003 Plant and animal 644 rhabdovirus host range: a bug's view. Trends in Microbiology 11(6), 264-271. (doi:Doi 645 10.1016/S0966-842x(03)00120-3). bioRxiv preprint doi: https://doi.org/10.1101/020107; this version posted August 6, 2015. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. 17

646 8. Ahne W., Bjorklund H.V., Essbauer S., Fijan N., Kurath G., Winton J.R. 2002 Spring 647 viremia of carp (SVC). Diseases of Aquatic Organisms 52, 261-272. 648 9. Pfeilputzien C. 1978 EXPERIMENTAL TRANSMISSION OF SPRING VIREMIA OF 649 CARP THROUGH CARP LICE (ARGULUS-FOLIACEUS). Zentralblatt Fur Veterinarmedizin 650 Reihe B-Journal of Veterinary Medicine Series B-Infectious Diseases Immunology Food 651 Hygiene Veterinary Public Health 25(4), 319-323. 652 10. Longdon B., Jiggins F.M. 2012 Vertically transmitted viral endosymbionts of 653 insects: do sigma viruses walk alone? Proc Biol Sci 279(1744), 3889-3898. 654 (doi:10.1098/rspb.2012.1208). 655 11. Longdon B., Wilfert L., Obbard D.J., Jiggins F.M. 2011 Rhabdoviruses in two 656 species of Drosophila: vertical transmission and a recent sweep. Genetics 188(1), 141- 657 150. (doi:10.1534/genetics.111.127696 ). 658 12. Longdon B., Wilfert L., Osei-Poku J., Cagney H., Obbard D.J., Jiggins F.M. 2011 659 Host switching by a vertically-transmitted rhabdovirus in Drosophila. Biology Letters 660 7(5), 747-750. (doi:10.1098/rsbl.2011.0160). 661 13. Li C.X., Shi M., Tian J.H., Lin X.D., Kang Y.J., Chen L.J., Qin X.C., Xu J., Holmes E.C., 662 Zhang Y.Z. 2015 Unprecedented genomic diversity of RNA viruses in arthropods reveals 663 the ancestry of negative-sense RNA viruses. eLife 4. (doi:10.7554/eLife.05378). 664 14. Walker P.J., Firth C., Widen S.G., Blasdell K.R., Guzman H., Wood T.G., Paradkar 665 P.N., Holmes E.C., Tesh R.B., Vasilakis N. 2015 Evolution of genome size and complexity 666 in the rhabdoviridae. PLoS Pathog 11(2), e1004664. 667 (doi:10.1371/journal.ppat.1004664). 668 15. Ballinger M.J., Bruenn J.A., Taylor D.J. 2012 Phylogeny, integration and 669 expression of sigma virus-like genes in Drosophila. Mol Phylogenet Evol 65(1), 251-258. 670 (doi:10.1016/j.ympev.2012.06.008). 671 16. Fort P., Albertini A., Van-Hua A., Berthomieu A., Roche S., Delsuc F., Pasteur N., 672 Capy P., Gaudin Y., Weill M. 2011 Fossil Rhabdoviral Sequences Integrated into 673 Arthropod Genomes: Ontogeny, Evolution, and Potential Functionality. Mol Biol Evol. 674 (doi:10.1093/molbev/msr226). 675 17. Katzourakis A., Gifford R.J. 2010 Endogenous Viral Elements in Animal Genomes. 676 Plos Genet 6(11), e1001191. (doi:10.1371/journal.pgen.1001191). 677 18. Aiewsakun P., Katzourakis A. 2015 Endogenous viruses: Connecting recent and 678 ancient viral evolution. Virology 479-480C, 26-37. (doi:10.1016/j.virol.2015.02.011). 679 19. Longdon B., Wilfert L., Jiggins F.M. 2012 The Sigma Viruses of Drosophila, Caister 680 Academic Press. 681 20. Rosen L. 1980 Carbon-dioxide sensitivity in mosquitos infected with sigma, 682 vesicular stomatitis, and other rhabdoviruses. Science 207(4434), 989-991. 683 21. Shroyer D.A., Rosen L. 1983 Extrachromosomal-inheritance of carbon-dioxide 684 sensitivity in the mosquito culex-quinquefasciatus. Genetics 104(4), 649-659. 685 22. Parker D.J., Vesala L., Ritchie M.G., Laiho A., Hoikkala A., Kankare M. 2015 How 686 consistent are the transcriptome changes associated with cold acclimation in two 687 species of the Drosophila virilis group? Heredity (Edinb). (doi:10.1038/hdy.2015.6). 688 23. van Mierlo J.T., Overheul G.J., Obadia B., van Cleef K.W., Webster C.L., Saleh M.C., 689 Obbard D.J., van Rij R.P. 2014 Novel Drosophila viruses encode host-specific suppressors 690 of RNAi. PLoS Pathog 10(7), e1004256. (doi:10.1371/journal.ppat.1004256). 691 24. Katoh K., Standley D.M. 2013 MAFFT multiple sequence alignment software 692 version 7: improvements in performance and usability. Mol Biol Evol 30(4), 772-780. 693 (doi:10.1093/molbev/mst010). 694 25. Capella-Gutierrez S., Silla-Martinez J.M., Gabaldon T. 2009 trimAl: a tool for 695 automated alignment trimming in large-scale phylogenetic analyses. Bioinformatics 696 25(15), 1972-1973. (doi:10.1093/bioinformatics/btp348). 697 26. Talavera G., Castresana J. 2007 Improvement of phylogenies after removing 698 divergent and ambiguously aligned blocks from protein sequence alignments. Syst Biol 699 56(4), 564-577. (doi:Doi 10.1080/10635150701472164). bioRxiv preprint doi: https://doi.org/10.1101/020107; this version posted August 6, 2015. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. 18

700 27. Guindon S., Dufayard J.F., Lefort V., Anisimova M., Hordijk W., Gascuel O. 2010 701 New Algorithms and Methods to Estimate Maximum-Likelihood Phylogenies: Assessing 702 the Performance of PhyML 3.0. Syst Biol 59(3), 307-321. (doi:Doi 703 10.1093/Sysbio/Syq010). 704 28. Le S.Q., Gascuel O. 2008 An improved general amino acid replacement matrix. 705 Molecular Biology and Evolution 25(7), 1307-1320. (doi:Doi 10.1093/Molbev/Msn067). 706 29. Anisimova M., Gascuel O. 2006 Approximate likelihood-ratio test for branches: A 707 fast, accurate, and powerful alternative. Syst Biol 55(4), 539-552. 708 (doi:10.1080/10635150600755453). 709 30. Rambaut A. 2011 FigTree. (v1.3 ed. 710 31. Hudson R.R., Slatkin M., Maddison W.P. 1992 Estimation of levels of gene flow 711 from DNA sequence data. Genetics 132(2), 583-589. 712 32. Bhatia G., Patterson N., Sankararaman S., Price A.L. 2013 Estimating and 713 interpreting FST: the impact of rare variants. Genome Res 23(9), 1514-1521. 714 (doi:10.1101/gr.154831.113). 715 33. Hommola K., Smith J.E., Qiu Y., Gilks W.R. 2009 A permutation test of host- 716 parasite cospeciation. Mol Biol Evol 26(7), 1457-1468. (doi:10.1093/molbev/msp062). 717 34. Jeyaprakash A., Hoy M.A. 2009 First divergence time estimate of spiders, 718 scorpions, mites and ticks (subphylum: Chelicerata) inferred from mitochondrial 719 phylogeny. Experimental & applied acarology 47(1), 1-18. (doi:10.1007/s10493-008- 720 9203-5). 721 35. Misof B., Liu S., Meusemann K., Peters R.S., Donath A., Mayer C., Frandsen P.B., 722 Ware J., Flouri T., Beutel R.G., et al. 2014 Phylogenomics resolves the timing and pattern 723 of insect evolution. Science 346(6210), 763-767. (doi:10.1126/science.1257570). 724 36. Bootsma R., Dekinkelin P., Leberre M. 1975 Transmission Experiments with Pike 725 Fry (Esox-Lucius L) Rhabdovirus. J Fish Biol 7(2), 269-276. (doi:Doi 10.1111/J.1095- 726 8649.1975.Tb04599.X). 727 37. Dorson M., Dekinkelin P., Torchy C., Monge D. 1987 SUSCEPTIBILITY OF PIKE 728 (ESOX-LUCIUS) TO DIFFERENT SALMONID VIRUSES (IPN, VHS, IHN) AND TO THE 729 PERCH RHABDOVIRUS. BULLETIN FRANCAIS DE LA PECHE ET DE LA PISCICULTURE 730 (307), 91-101. 731 38. Haenen O., Davidse A. 1993 Comparative pathogenicity of two strains of pike fry 732 rhabdovirus and spring viremia of carp virus for young roach, common carp, grass carp 733 and rainbow trout. Diseases of Aquatic Organisms 15(2), 87-92. 734 39. Okland A.L., Nylund A., Overgard A.C., Blindheim S., Watanabe K., Grotmol S., 735 Arnesen C.E., Plarre H. 2014 Genomic characterization and phylogenetic position of two 736 new species in Rhabdoviridae infecting the parasitic copepod, salmon louse 737 (Lepeophtheirus salmonis). Plos One 9(11), e112517. 738 (doi:10.1371/journal.pone.0112517). 739 40. Drummond A.J., Suchard M.A., Xie D., Rambaut A. 2012 Bayesian phylogenetics 740 with BEAUti and the BEAST 1.7. Mol Biol Evol 29(8), 1969-1973. 741 (doi:10.1093/molbev/mss075). 742 41. Weinert L.A., Welch J.J., Suchard M.A., Lemey P., Rambaut A., Fitzgerald J.R. 2012 743 Molecular dating of human-to-bovid host jumps by Staphylococcus aureus reveals an 744 association with the spread of domestication. Biol Lett 8(5), 829-832. 745 (doi:10.1098/rsbl.2012.0290). 746 42. Drummond A.J., Ho S.Y., Phillips M.J., Rambaut A. 2006 Relaxed phylogenetics 747 and dating with confidence. PLoS Biol 4(5), e88. (doi:10.1371/journal.pbio.0040088). 748 43. Minin V.N., Suchard M.A. 2008 Counting labeled transitions in continuous-time 749 Markov models of evolution. Journal of mathematical biology 56(3), 391-412. 750 (doi:10.1007/s00285-007-0120-8). 751 44. Rambaut A., Drummond A.J. 2007. Tracer v16, Available from 752 http://beast.bio.ed.ac.uk/Tracer bioRxiv preprint doi: https://doi.org/10.1101/020107; this version posted August 6, 2015. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. 19

753 45. Longdon B., Obbard D.J., Jiggins F.M. 2010 Sigma viruses from three species of 754 Drosophila form a major new clade in the rhabdovirus phylogeny. Proceedings of the 755 Royal Society B 277, 35-44. (doi:10.1098/rspb.2009.1472). 756 46. Longdon B., Walker P.J. 2011 ICTV sigmavirus species and genus proposal. ( 757 47. Walker P.J., Dietzgen R.G., Joubert D.A., Blasdell K.R. 2011 Rhabdovirus accessory 758 genes. Virus Res 162(1-2), 110-125. (doi:10.1016/j.virusres.2011.09.004). 759 48. Chiba S., Kondo H., Tani A., Saisho D., Sakamoto W., Kanematsu S., Suzuki N. 2011 760 Widespread Endogenization of Genome Sequences of Non-Retroviral RNA Viruses into 761 Plant Genomes. Plos Pathogens 7(7). (doi:Artn E1002146 762 Doi 10.1371/Journal.Ppat.1002146). 763 49. Aznar-Lopez C., Vazquez-Moron S., Marston D.A., Juste J., Ibanez C., Berciano J.M., 764 Salsamendi E., Aihartza J., Banyard A.C., McElhinney L., et al. 2013 Detection of 765 rhabdovirus viral RNA in oropharyngeal swabs and ectoparasites of Spanish bats. J Gen 766 Virol 94(Pt 1), 69-75. (doi:10.1099/vir.0.046490-0). 767 50. L'Heritier P.H., Teissier G. 1937 Une anomalie physiologique héréditaire chez la 768 Drosophile. CR Acad Sci Paris 231, 192-194. 769 51. Longdon B., Brockhurst M.A., Russell C.A., Welch J.J., Jiggins F.M. 2014 The 770 Evolution and Genetics of Virus Host Shifts. PLoS Pathog 10(11), e1004395. 771 (doi:10.1371/journal.ppat.1004395). 772 52. Longdon B., Hadfield J.D., Webster C.L., Obbard D.J., Jiggins F.M. 2011 Host 773 phylogeny determines viral persistence and replication in novel hosts. PLoS Pathogens 774 7((9)), e1002260. (doi:10.1371/journal.ppat.1002260). 775 53. Faria N.R., Suchard M.A., Rambaut A., Streicker D.G., Lemey P. 2013 776 Simultaneously reconstructing viral cross-species transmission history and identifying 777 the underlying constraints. Philos Trans R Soc Lond B Biol Sci 368(1614), 20120196. 778 (doi:10.1098/rstb.2012.0196). 779 54. Streicker D.G., Turmelle A.S., Vonhof M.J., Kuzmin I.V., McCracken G.F., Rupprecht 780 C.E. 2010 Host Phylogeny Constrains Cross-Species Emergence and Establishment of 781 Rabies Virus in Bats. Science 329(5992), 676-679. (doi:10.1126/science.1188836). 782 55. Longdon B., Hadfield J.D., Day J.P., Smith S.C., McGonigle J.E., Cogni R., Cao C., 783 Jiggins F.M. 2015 The Causes and Consequences of Changes in Virulence following 784 Pathogen Host Shifts. PLoS Pathog 11(3), e1004728. 785 (doi:10.1371/journal.ppat.1004728). 786 56. Wang Z., Gerstein M., Snyder M. 2009 RNA-Seq: a revolutionary tool for 787 transcriptomics. Nat Rev Genet 10(1), 57-63. (doi:10.1038/nrg2484). 788 57. Dudas G., Obbard D.J. 2015 Are arthropods at the heart of virus evolution? eLife 789 4. 790 58. Aguiar E.R., Olmo R.P., Paro S., Ferreira F.V., de Faria I.J., Todjro Y.M., Lobo F.P., 791 Kroon E.G., Meignin C., Gatherer D., et al. 2015 Sequence-independent characterization 792 of viruses based on the pattern of viral small RNAs produced by the host. Nucleic Acids 793 Res. (doi:10.1093/nar/gkv587). 794 59. Webster C.L., Waldron F.M., Robertson S., Crowson D., Ferrai G., Quintana J.F., 795 Brouqui J.M., Bayne E.H., Longdon B., Buck A.H., et al. 2015 The discovery, distribution 796 and evolution of viruses associated with Drosophila melanogaster. PLOS Biology 13(7): 797 e1002210. 798 799 800 All data has been made available in public repositories: 801 802 NCBI Sequence Read Archive Data: SRP057824 803 Data S1, sample information: http://dx.doi.org/10.6084/m9.figshare.1425432 804 Data S2, virus ID, Genbank accessions and host information: 805 http://dx.doi.org/10.6084/m9.figshare.1425419 bioRxiv preprint doi: https://doi.org/10.1101/020107; this version posted August 6, 2015. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC 4.0 International license. 20

806 Figures S1-S3: http://dx.doi.org/10.6084/m9.figshare.1495351 807 L gene sequences fasta: http://dx.doi.org/10.6084/m9.figshare.1425067 808 TrimAl alignment fasta: http://dx.doi.org/10.6084/m9.figshare.1425069 809 Gblocks alignment fasta: http://dx.doi.org/10.6084/m9.figshare.1425068 810 Phylogenetic tree Gblocks alignment: http://dx.doi.org/10.6084/m9.figshare.1425083 811 Phylogenetic tree TrimAl alignment: http://dx.doi.org/10.6084/m9.figshare.1425082 812 BEAST alignment fasta: http://dx.doi.org/10.6084/m9.figshare.1425431 813 BEAUti xml file: http://dx.doi.org/10.6084/m9.figshare.1431922 814 Bayesian analysis tree: http://dx.doi.org/10.6084/m9.figshare.1425436 815 Tables S1-5. List of newly discovered viruses: 816 http://dx.doi.org/10.6084/m9.figshare.1502665 817 818